Hi, Bradley.
I tried to simulate the echoes of an air or water filled ball or SDH in a steel layer. The problem is the echo reflected from the upper boundary is so strong that the ball echo was masked. However, when there is a steel ball in water, it is easy to find the echo and get a good beamformed image.
The frequency is 5MHz, and the grid points per wavelength is 4 at 5MHz.
What is the problem? Thanks!
Best Regards
Biao
Hi Biao,
I'm not sure I completely follow your setup. What is the upper boundary that you get a strong reflection from? Is it the steel layer? Is this also in water?
Brad.
Hi, Bradley,
Yes, there is a steel layer immersed in water, and also the transducer array is in the water, the objective is to inspect the flaws inside the steel material layer.Since there is a big difference between the acoustic the impedence of the steel and the water, most of the energy, say more than 90% are reflected by the upper boundary of the steel layer. As a result, if there are air voids or other cracks in the layer, we can't find the flaw echo.
So, since there are solid layer in liquid, whether we must use function pstdElastic3D instead of kspaceFirstOrder3D?
Can pstdElastic3D work well when exists large medium inhomogeneity? If not,the only thing we can do is to scale the material parameters.
Thanks!
Best Regards
Biao
Hi Biao,
If the physical situation has a large impedance mismatch, then there's not a lot you can do unfortunately! If you gate out the reflection from the front surface, or subtract the recorded signal when there is no flaw, can you see flaw echo then?
If you want to model the contribution of shear waves in the solid layer, then it's probably best to use pstdElastic3D. You could always compare the output between this and kspaceFirstOrder3D to see if this is important for your particular configuration.
To get accurate reflection coefficients for very large impedance changes, you will need to use a sufficient number of points per wavelength (see here).
Hope that helps,
Brad.
Hi, Bradley,
Thanks, I increased the grid points per wavelength when existing larger impedance mismatch, but the computation time increased much, especially for higher frequency, For 5MHz, 8-transducer array, and each transducer acted as a transmitter sequentially, it ran for one whole day to get the 3-D A-scan data, and the synthetic aperture imaging results were nice, we can detect the defects inside a thin steel layered material.
Best Regards
Biao
You must log in to post.